What do Americans Think About the Police?

Killings by law enforcement in the US, while unfortunate, have also played an integral role in shaping the way Americans feel towards and experience the modern day justice system. I will leave the latter to the experts, pointing to some studies and reviews later on for those who are curious about learning more about the determinants of civilian killings.

What I can say is that there exists a great disparity among Americans with regards to the attitudes they hold towards the police and police brutality. A recent poll performed by the Associated Press-NORC Center for Public Affairs Research compared sentiment towards police violence after Freddie Gray’s death in 2015 and George Floyd’s death in the present day. Almost half of Americans say police violence against the public is extremely or very serious today, up from 2015 when this view was shared by only about a third of all adults.

When taking race into account, however, about 40% of White adults agree that it is an extremely or very serious problem, compared to the 83% of Black adults who share the same view. What can be agreed upon, regardless of race, is that police violence against the public is an important issue - with 79% of Americans thinking it is at least moderately serious. Here are other things Americans agree on, according to this poll:

Policy All Americans Black Americans White Americans
Requiring on-duty police officers to wear video cameras that would record their interactions with the public as they occur 88% 90% 91%
Requiring officers to report misconduct by their peers 87% 87% 88%
Reducing funding for law enforcement agencies 25% 43% 22%
The criminal justice system needs at least major changes 69% 92% 65%
The criminal justice system treat police who injure or kill civilians too leniently 65% 84% 62%

The first three are of the percentage who strongly/somewhat favor policies to prevent police. These results seem to suggest that while some policies are favorable, others are not as supported - such as reducing funding for law enforcement agencies. The survey methodology for this poll can be found here

While we’ll probably see more data start to bubble up for what has happened this year, we can still learn more from the surveys conducted in the wake of the 2015 Black Lives Matter movement. For this, we can turn to the Cato Institute’s “Policing in America: Understanding Public Attitudes Toward the Police,” whose survey methodology can be found here.

Looking at how favorably the police has been viewed by Americans over time, we can see that very little has changed. In 50 years, the percentage of Black and White Americans who have a favorable view of the police has increased by only 3 and 1 percentage points, respectively.

Race 1970 2015
All Americans 64%
Black Americans 43% 40%
White Americans 67% 68%
Hispanic Americans 59%

According to the survey, feelings towards police aren’t just variable across race, but also age, income and geography. The younger generations and those with lower income tend to have a less favorable outlook towards local police. In terms of geographic location, people living in suburbs have a slightly better view of local police.

Age Favorable Toward Local Police Definitely Would Report a Crime
18-29 54% 54%
30-44 54% 63%
45-64 70% 82%%
65+ 82% 87%
Income Favorable Toward Local Police Definitely Would Report a Crime
<$30,000 54% 62%
$30,000-49,999 64% 73%
$50,000-99,999 76% 76%
100,000+ 76% 84%
Location Favorable Toward Local Police Definitely Would Report a Crime
City 60% 69%
Suburb 69% 74%
Rural 61% 72

It will be interesting to see how these numbers evolve in response to the latest wave of high-profile police killings involving people like Breonna Taylor, George Floyd, Ahmaud Arbery, Tony McDade, Dion Johnson and more. For more information on the data set used to make the following visualizations, as well as the media bias rating for the organizations mentioned previously, take a look at the notes at the end.

13 Visualizations of Police Killings in America from 2000 - 2016

The following visualizations are intended to help you get a broader picture of the makeup of police killings from 2000 to 2016. Only cases for which all information was available were included, excluding, for example, whether or not the person was armed because of the large number of missing values.

Police Killings by Year

Police killings in 2000 were 316, practically doubling a decade later in 2010 to 625. From the time period studied here, 2015 was the deadliest year yet over the 16-year period with 1573 deaths. This number was 1,146 according to the Guardian and about 1,000 according to the Washington Post.

The number of police killings have been increasing when controlling for the increasing size of the population. At the beginning of the decade, 2 people were killed per million in the population. Just six years later, this number increased by 65% to about 3.3.

Looking at police killings per quarter, there is a drop off after the first quarter of 2016. Keep in mind the whole month of December does not have any observations for the fourth quarter of 2016.

Police Killings by Age Group

The age groups with the largest share of fatalities are those aged 31-40 and 19-25, who account for 26.2% and 22% of killings from 2000 - 2016, respectively. People aged 19 to 50 account for 82% of police killings for the 16-year period.

Conditions Surrounding Deaths

Overwhelmingly, the manner of death for those killed by police is being shot. For every year of this 16-year period, fatal shootings account for 90% or more of police killings.

From 2000 to 2016, the mental state of people killed has been about 20% for all years.

The lack of “yes”’s between 2000 - 2014 could be due to fault of reporting. Taking the data as it is, however, seems to suggest that all if not the majority of police killings occur with people not attempting to flee.

Police Killings by Race

One quick note about the race variable. The data set used only indicated “Native” as a race, which is most likely how it is reported in the agencies’ system. Race and ethnicity have always been hard to record properly. Not wanting to cherry pick population sizes, I included Native Americans, Alaska Natives, Native Hawaiians and other Pacific Islanders under a broad category.

An imperfect solution, I agree, however I urge you to look at some other databases. In 2016, for example, the Guardian reported Native Americans had about 10 police killings per million in the population.

Looking at police killings by race, White Americans hold the biggest share for the majority of years from 2000 to 2016. This is expected, as White Americans make up the majority of the population in the US.

The decline in missing values (NA) suggests an improvement in reporting.

Looking at police killings by race per half million in the population of each race, Black Americans have been killed disproportionately throughout the 16-year period, followed by Hispanic Americans and native Americans.

Police Killings by Gender

It is abundantly clear that males are disproportionately effected by police killings. For 2016, females made up about 4.3% of police killings while males accounted for about 95.5%. About 0.2% of police killings in 2016 did not have a gender specified.

Mapping Out Police Violence

Looking at the number of police killings in each state, the highest number of cumulative killings between 2000 - 2016 took place, in descending order, in:

  1. California
  2. Texas
  3. Florida
  4. Arizona
  5. New York

Looking instead to police killings by the population of each state, the highest number of cumulative police killings per 100,000 in the population for 2000 to 2016 were, in descending order:

  1. Alaska
  2. Nevada
  3. District of Columbia
  4. Arizona
  5. California

Out of 4,000 cities, there were about 41 whose coordinates could not be found by the geocode_OSM function used to find them.

The cities with the highest number of police killings throughout this 16-year period were, in descending order:

  1. Los Angeles, California
  2. Houston, Texas
  3. Chicago, Illinois
  4. Phoenix, Arizona
  5. Las Vegas, Nevada

Populations of all cities were not readily available, or else killings per thousand in each city could have been calculated.

Notes

As I mentioned, the way Americans experience the justice system is a more nuanced subject. While you may see a lot of people jumping to conclusions based on survey results, their have been many detailed studies that have been conducted on police brutality that paint a clearer picture. I suggest three things:

The first two may seem paradoxical, but in many cases are not. Personal testimonies, while often viewed as subjective, can frequently lead to an uncovering of objective facts. One example of this can be seen in the testimonies of survivors of police torture in the 1980s. This group of Chicago police officers, who became known as the “midnight crew,” were at large for 20 years and tortured over 100 black men and women. The personal testimonies of Darrell Cannon and Ronald Kitchen helped jump-start the investigation into the police officers who perpetrated the heinous crimes against them.

In that vein, experts can often offer an objective view into various subject matters, specifically those that have specializations in the field. Here are some studies you can use to start your research:

Lastly, keeping media bias in mind has become vital in all aspects of news consumption. Here are the media ratings on the sources quoted in this article based on two organizations.

Entity Media Bias Fact Check Allsides
Associated Press (AP) Least biased Center
Cato Institute Right-center Lean Right
The Appeal Center
Los Angeles Times Left-center Lean Left
The Guardian Left-center Lean Left
The Washington Post Left-center Lean Left

If you’re curious about the data set - good! As far as data analysis goes, you’re analysis is only as good as your data. Unfortunately, data on police killings are not provided by the government (as a lot of important statistics aren’t). However, there are a couple of data sets and databases you can check out if you’re interested in this subject:

I chose the second data set, as the first only had data for the year 2015. The FBI, which is the body responsible for gathering crime statistics, launched a data collection program in 2015 in order to address the severe lack of data on use-of-force by police in the field. Unfortunately, law enforcement agencies are only encouraged - not required - to participate.

In contrast, statistics on law enforcement officers killed in the line of duty are collected yearly by the FBI.

While unsatisfactory or absent data is a common problem, statistics as vital as these should not be missing. Collecting data on police killings committed off-duty or non-fatal uses of force could also give us a more nuanced view of police brutality, as argued by the LA Times.

The data set from Data World had a lot of cleaning up to do. Some major examples:

If you’re interested in understanding more about the data-wrangling process behind this analysis, feel free to drop me a message.